Looking to develop a system (web or non-web-based) that allows me to do the following:
1. Web or non-web-based code that scans digitized book content and and based on keywords, makes judgments about how that book should be categorized. Electronically both review and categorise a backlist of 5,000+ books (novels). They will to be categorized by various criteria including timeframe, location, content, style, events, and other parameters/attributes I would provide.
2. take data on the previous book purchase history of a wide range of users and, combine with the categorisation work done in 1. to build profiles of users to demonstrate their likes and dislikes;
3. suggest future book purchases based on profiles generated. Format sales emails etc with ability to purchase titles through API.
Basically, I want a database of books to be more fully categorized by additional attributes and want code to be developed to apply this new categorization scheme to each books to generate better recommendations to users based on their book purchase history.
Regarding your first question, the search challenge here will be (it seems to me) to be taking an entire novel and categorising it. If the book genre were action novels (which it is not) the system would be required to determine without additional human input (ie beyond providing the categories):
1. time period (WW1, WW2, Vietnam, Ancient Rome etc);
2. type of action (naval, aerial, hand-to-hand, mass, secret agent etc);
3. number and intensity of battle scenes (even goriness, potentially);
4. location (global, Europe, Asia etc);
5. style (gritty, lyrical, classical, earthy)
6. anything else you care to think of.
in addition the programme would note price point, whether purchased in response to a price promotion, its author, date of purchase etc (all of which could be provided).
Open to open-source solutions. Might involve use of [url removed, login to view] or Autonomy IDOL or anything else you could recommend that would address some or all of the requirements above.
Please read the project first and write the word "Understood". Also write your steps and suggestions to complete the project with a short description of what you understood.